eHealth 360°. International Summit on eHealth, Budapest, Hungary, June 14-16, 2016, Revised Selected Papers

Research Article

Non Invasive Detection of Coronary Artery Disease Using PCG and PPG

Download
471 downloads
  • @INPROCEEDINGS{10.1007/978-3-319-49655-9_32,
        author={Rohan Banerjee and Anirban Choudhury and Shreyasi Datta and Arpan Pal and Kayapanda Mandana},
        title={Non Invasive Detection of Coronary Artery Disease Using PCG and PPG},
        proceedings={eHealth 360°. International Summit on eHealth, Budapest, Hungary, June 14-16, 2016, Revised Selected Papers},
        proceedings_a={EHEALTH360},
        year={2017},
        month={1},
        keywords={Coronary Artery Disease Phonocardiogram Classification Photoplethysmogram Transfer function},
        doi={10.1007/978-3-319-49655-9_32}
    }
    
  • Rohan Banerjee
    Anirban Choudhury
    Shreyasi Datta
    Arpan Pal
    Kayapanda Mandana
    Year: 2017
    Non Invasive Detection of Coronary Artery Disease Using PCG and PPG
    EHEALTH360
    Springer
    DOI: 10.1007/978-3-319-49655-9_32
Rohan Banerjee1,*, Anirban Choudhury1,*, Shreyasi Datta1,*, Arpan Pal1,*, Kayapanda Mandana2,*
  • 1: Tata Consultancy Services Ltd.
  • 2: Fortis Hospital
*Contact email: rohan.banerjee@tcs.com, anirban.duttachoudhury@tcs.com, shreyasi.datta@tcs.com, arpan.pal@tcs.com, kmmandana@gmail.com

Abstract

Coronary Artery Disease (CAD) kills more than a million of people every year. However, there is no significant marker for identifying CAD patients unobtrusively. In this paper, we propose a methodology for non invasive screening of CAD patients from heart sound analysis. Instead of segregating the diastolic heart sound as mentioned in prior arts, the proposed methodology extracts spectral features from the entire phonocardiogram (PCG) signal, broken into small overlapping windows. Support vector machine (SVM) is used for classification. Our methodology produces 80% classification accuracy on a dataset of 25 subjects, containing PCG data of both cardiac an non cardiac patients as well as healthy subjects. Results also reveal that a simple transfer function can be formed to identify the CAD patients if photoplethysmogram (PPG) signal is available simultaneously along with PCG.